Although there are several effective treatments for a major depressive episode, there are no reliable predictors of the likelihood of remission, response or non-response with an initial trial of either an antidepressant medication or psychotherapy. In prioritizing a role for direct measures of brain functioning in the development of new algorithms for clinical management of depressed patients, a systematic characterization of pretreatment patterns predictive of unambiguous remission to standard treatments is a necessary first step. This project will characterize imagingbased brain subtypes that distinguish groups of never-treated depressed patients who subsequently respond to pharmacotherapy or cognitive behavior therapy (CBT), respectively. A prospectively-treated cohort of 400 never-treated depressed patients randomized to receive either escitalopram, duloxetine or CBT for 1.2 weeks will define these subtypes. Resting-state BOLD functional magnetic resonance imaging (fMRI) scans will be acquired prior to initiating antidepressant therapy and at a fixed, early time point specific for each treatment. Pre-treatment scan patterns derived using multivariate analyses and associated with the six possible response outcomes (3 types of response;3 types of nonresponse) will be used to determine whether pretreatment brain patterns can distinguish among outcome groups. A second fMRI scan, acquired early in the treatment course, will be used to assess the likelihood of response to the specific treatment assigned. The proposed studies are a first step towards defining brain-based subtypes predictive of differential treatment outcome in major depression. The data from these studies will also be entered into more complex algorithms integrating imaging findings with behavioral, environmental, biochemical and genetic information for individual patients.